Bug #74564 ALLOC_NODEID_REQ succeeds when it shouldn\'t

• Submitted: 25 Oct 2014 14:56
• Modified: 22 Dec 2014 19:06
• Reporter: Mikael Ronström
• Status: Closed
• Category: MySQL Cluster: Cluster (NDB) storage engine
• Severity: S3 (Non-critical)
• Version: 7.4.2
• OS: Any
• CPU Architecture: Any

[25 Oct 2014 14:56] Mikael Ronström

Description:
When a node is restarting from the angel process it starts by allocating a new nodeid.
This should not succeed until we have completed the node failure handling. However it
now 

How to repeat:
Found in 7.4 when introducing more stringent state machine that checks that we transition from state to state in proper ways in
the node restart handling. It is not known to cause any major problems yet. It is however not a good idea to have a starting node
run while we are processing the node failure, so a bug can easily be there or easily be introduced.

Suggested fix:
Ensure that we also check if the node is in cfailedNodes in QMGR where it resides during the commit of the node failure.
Also handle timeout of the node with the id we are looking for in a special manner.

[22 Dec 2014 19:06] Jon Stephens

Thank you for your bug report. This issue has already been fixed in the latest released version of that product, which you can download at

  http://www.mysql.com/downloads/

Documented fix in the NDB 7.4.3 changelog, as follows:

    When a data node was restarted from its angel process (that is,
    following a node failure), it could be allocated a new node ID
    before failure handling was actually completed for the failed
    node.
      
Closed.